Elucidating the MARine DIversity GRAdient with empirical and theoretical models

Project Information

MARDIGRA

Grant agreement ID: 891702

DOI

10.3030/891702

Project closed

EC signature date 15 April 2020

Start date 1 May 2020

End date 30 April 2022

Funded under

• EXCELLENT SCIENCE - Marie Skłodowska-Curie Actions

Total cost

€ 207 312,00

EU contribution

€ 207 312,00

Coordinated by

KOBENHAVNS UNIVERSITET
Denmark

Periodic Reporting for period 1 - MARDIGRA (Elucidating the MARine DIversity GRAdient with empirical and theoretical models)

Reporting period: 2020-05-01 to 2022-04-30

Summary of the context and overall objectives of the project

As Earth warms, our oceans are predicted to undergo dramatic changes. However, anticipating the effects of climate change on life in our oceans will remain a difficult task without understanding what marine biodiversity patterns are and how they were generated. To date, study research on distributions and drivers of global biodiversity has been dominated by work in terrestrial systems, and uncertainty persists regarding whether terrestrial and marine biodiversity are shaped by analogous processes. MARDIGRA sought to establish a new analytical procedure for estimating baseline marine diversity patterns and inferring the broad-scale processes that generated those patterns.

The specific objectives of MARDIGRA were to 1) pioneer a workflow for modeling species distributions of three orders of marine fishes in the Atlantic, 2) model how environmental conditions and past climate changes may have shaped modern marine diversity patterns, and 3) compare model results for the three fish orders to determine the degree to which the effects of climate on diversity are universal. I chose three well-known fish orders for this study (Gadiformes, codfishes and allies; Scombriformes, tunas and allies; and Beloniformes, flyingfishes and allies) because they generally inhabit three different marine depth regions—bottom, midwaters, and surface waters, respectively. Furthermore, these orders contain many commercially important species under intensive fishing pressure. MARDIGRA strengthens links between observed geographic concentrations of species diversity and theory-predicted expectations of how historical climate change shaped these patterns and illuminates unique processes underlying marine diversity patterns. In turn, project findings will aid in anticipating potential effects of climate change on future diversity, and lead to more robust preservation, fisheries management, and climate change mitigation strategies.

Work performed from the beginning of the project to the end of the period covered by the report and main results achieved so far

Work Package 1: Empirical estimates of fish diversity patterns. I collected occurrence data for all the fishes of interest (434 species) from online databases and the literature. Next, I conceived of a novel, 3-D approach to infer species distributions from occurrences and bundled the resulting software tools into a new R package, voluModel. Species distribution modelling and related analytical methods were first developed in two-dimensional terrestrial systems; many common workflows organize and analyze geographically structured occurrence and environmental data based on horizontal latitude and longitude coordinates. However, marine species also distributed vertically, and water conditions may vary strikingly with depth. voluModel's innovative tools facilitate generation of fine-grained, 3-D distribution models to infer potential species distributions. I then modeled species distributions based on environmental conditions where each species was observed, using both standard methods to model horizontal distributions, and my new method to model 3-D distributions.

Work package 2: Process-based models of marine biodiversity patterns. WP2 has mostly consisted of gathering and processing mapped past and present environmental data for ecological variables. According to theory, response to both oceanographic conditions and how stable these conditions were may have driven the emergence of marine biodiversity patterns. I downloaded decadal mean annual average ocean salinity, temperature, and ideal water age (i.e. time since surface contact) between 22,000 and 400 BCE from climate simulations run using two different climate models and I am converting these data into 100-year climatologies at 1,000 year time steps over the last 20,000 years. I will then calculate Holocene and Late Pleistocene temperature, salinity, and water age stability. I have also obtained modern oceanographic data on water density and nutrients.

Work package 3: Describe effects of various drivers on marine biodiversity patterns. Upon completion of species distribution models (WP1), I estimated Atlantic Ocean species richness for each of the three orders of interest by stacking both 2-D and 3-D species distributions. While 2-D species distributions recreated biodiversity patterns like those of other studies using 2-D distributions, patterns of biodiversity based on 3-D models differed substantially, especially for Scombriformes and Gadiformes, which generally dwell below the surface. Once explanatory variable data processing is complete (WP2), these data will be used to examine which variables are most strongly correlated with observed biodiversity patterns, both in horizontal space and with depth, using a machine learning workflow I am currently developing.

Preliminary results of WP1 were disseminated as a live virtual oral presentation at the Joint Meeting of Ichthyologists and Herpetologists in July 2021. Preliminary results of WP1 and WP3 were presented virtually at the International Biogeography Society (IBS) meeting in Vancouver in June 2021 and received an award for best on-demand talk of the conference. The first MARDIGRA project manuscript, describing the voluModel R software package I created in WP1, is currently in review. The software is publicly archived and indexed in OpenAire. Two additional publications are planned. “The influence of natural history on biodiversity patterns: Lessons from fishes” is in preparation and will present empirical biodiversity patterns uncovered using the 3-D distribution modelling workflow. The final manuscript, presenting WP3 results, “Biodiversity gradients in marine systems: Uniting theory and empirical observation to understand processes behind patterns”, will use explanatory variables generated by WP2 to model WP1 observed empirical biodiversity patterns.

Progress beyond the state of the art and expected potential impact (including the socio-economic impact and the wider societal implications of the project so far)

To date, MARDIGRA's clearest advancement beyond the state of the art was the unanticipated breakthrough in how to model species distributions in 3-D space more efficiently and accurately, a long-standing need of marine studies. From a modern biodiversity perspective, stacking 3-D distributions will yield more precise estimates of open-ocean species richness and illuminate how biodiversity varies with depth. Projecting 3-D distributions through time will estimate more accurately how suitable conditions for marine species shifted both with latitude and depth in the past, as well as predicting how suitable habitat may shift in the future. Both improved biodiversity estimates and habitat suitability predictions can contribute substantially toward data-driven conservation efforts, including sustainable biodiversity management plans and robust marine protected area design.

These impacts will be further enhanced by the expected findings of WP 2 and 3, which will will demonstrate whether temperature, salinity, and water age stability since the Late Holocene and/or modern primary productivity are correlated with species richness for Gadiformes, Scombriformes, and Beloniformes. Given that species richness patterns are dissimilar among the groups, it will be possible to correlations among explanatory variables for the different orders of fishes, and assess which variables are most important for each. This is particularly valuable as it may provide insights into historical drivers of pelagic marine diversity and improve our capacity to anticipate the effects of ongoing climate change for disparate groups of economically-important fishes.

Visual summary of voluModel R functions and package contribution to efficient 3D ENM workflow.